Article by Eric Worrell
Taking “model output is data” to a new level…
Artificial intelligence reveals Europe’s hidden climate extremes
Andre Ionescu
Earth.com Staff Writer…
Traditionally, climate scientists have relied on statistical methods to interpret these data sets, but a recent breakthrough demonstrates the power of artificial intelligence (AI) to revolutionize this process.
Previously unrecorded climate extremes
A team led by Étienne Plésiat from the German Center for Climate Computing in Hamburg, together with colleagues from the UK and Spain, applied artificial intelligence to reconstruct climate extremes in Europe.
The study not only confirmed known climate trends but also revealed previously unrecorded extreme events.
…
The team used historical simulations from the CMIP6 archive (Coupled Model Intercomparison Project) to train CRAI to reconstruct past climate data.
Experts validated their results using standard metrics such as root mean square error and Spearman's rank correlation coefficient, which measure accuracy and correlation between variables.
…
Learn more: https://www.earth.com/news/ai-reveals-hidden-climate-extremes-in-europe/
The only real thing about using generative artificial intelligence to try to fill in the gaps is the illusion.
What is the Artificial Intelligence Illusion?
AI hallucination is a phenomenon in which a large language model (LLM) (usually a generative AI chatbot or computer vision tool) perceives patterns or objects that are not present or detectable by human observers, creating meaningless or completely inaccurate output.
Learn more: https://www.ibm.com/topics/ai-hallucinations
I am a fan of artificial intelligence and I believe that artificial intelligence is and will continue to make great contributions to the progress of mankind. But you have to test the output rigorously. Comparing AI output to a flawed model to see if it makes sense is not what I would call a test.
Climate scientists have been repeatedly criticized for treating model output as data. Using a tool prone to producing false or misleading data to produce a climate “record” that cannot be properly examined seems to me a science fiction – and a complete waste of time and money.
Relevant
